With over 200,000 new cases diagnosed yearly, it is imperative to understand breast cancers' genetic causes and characteristics. Current understanding of non-invasive breast cancer is limited, and knowledge of premalignant disease is rudimentary. Yet knowledge of genetic abnormalities seminal to early stages of carcinogenesis is critical for estimating cancer risk, diagnosing premalignant lesions destined to progress, selecting potential targets for new preventive agents, and generating a comprehensive understanding of cancer etiology. To address this critical gap in knowledge of breast carcinogenesis, this application's objective is to evaluate RNA expression in early lesions from breast cancer-containing tissues. We will use microdissection and oligonucleotide microarrays to examine RNA expression in co-existing epithelium (normal, atypical ductal hyperplastia (ADH) and ductal carcinoma in situ (DCIS)), and in stroma (fibroblasts and myoepithelium) surrounding each lesion. This project's rationale is the epidemiologic, histologic and genetic evidence that hyperplastic lesions may be non-obligate breast cancer precursors, and the growing data that the micro-environment plays a crucial role in tumor progression. However, the precise relationship between hyperplasias and cancers is not defined, and the role of the microenvironment in preinvasive disease is unexamined. Based on the existing evidence and our preliminary data, we speculate that distinct changes occur in the epithelium of these early lesions, and that the stroma surrounding normal and/or ADH lesions may also influence these lesions' progression. To accomplish our objective, we plan: Aim 1: Define RNA expression in preinvasive breast epithelium, a) obtain fresh tissue from 20 sporadic cancer cases, b) microdissect co-existing normal, ADH and DCIS lesions, extract and amplify RNA, perform array hybridizations, c) identify genes and pathways activated during progression using a novel bioinformatics method Bayesian Analysis of Differential Gene Expression (BADGE) and a traditional approach Significance Analysis of Microarrays (SAM) plus existing annotation resources. Aim 2: Define RNA expression in the stroma surrounding each epithelial lesion, a) from the same cases, isolate homogeneous cell populations in the surrounding stroma. b) extract and amplify RNA, perform array hybridizations, c) use BADGE, SAM and annotation resources to identify genes and pathways critical to the early stages of breast carcinogenesis.